Imagine the following conversation:
A: I was thinking the other day about Casement.
B: Oh yes.
A: I think the way he was treated was terrible.
B (looking puzzled): what do you mean? (suddenly understands) oh you mean Roger Casement the Irish Nationalist and poet? I thought you meant Patrick Casement the psychoanalyst.
So what was involved in B’s thinking that his friend was talking about Patrick Casement when he was really talking about Roger Casement? Presumably it was something that happened at the moment Casement was first mentioned – perhaps an image flashed briefly through B’s mind? But B doesn’t mention an image and although sometimes when someone mentions a name, an image does flash through our minds, it does not always happen. Furthermore, an image wouldn’t really get us very far, even if B did report one. We might ask: was it an image of a seated or a standing Patrick Casement? And if someone suggests that perception of the image was so fleeting that this question is inappropriate, we might well ask: so how could B be so sure the image was of Patrick Casement if he only saw if for a split second?
These Wittgensteinian reflections should leave us a bit puzzled about how we use language in this area – what is going on when we confidently explain what we meant? Or when we explain that we thought someone meant one thing when it later turned out they meant something else? The reference to the past seems to be essential – after all, we are interested in what the person thought/meant at the time because it is this that explains what they said next or how they reacted etc. My understanding of Wittgenstein is that it is misguided to look for something else to justify the individual’s claim that that was what he mean or what he thought the other person meant. There is no mental event distinct from the individual’s explanation.
Furthermore, in terms of brain states Wittgenstein would be highly sceptical about there being any change in the brain that corresponds to B thinking that A meant Patrick Casement when A actually meant Roger Casement. I don’t think Wittgenstein would be totally dogmatic on this – in the sense that if people do want to analyse what happens in people’s brains in these sort of situations, then they can of course do so and maybe some interesting results will emerge. However, one would have to be careful interpreting the results. Suppose we did develop a machine that in these situations was able to give 100% accurate information on what the listener took the speaker to mean. So in our example the machine on the basis of some kind of brain scan would indicate that Casement had initially been understood as Patrick Casement and then when we question B he would confirm this.
It is difficult to see quite how this might work really, but perhaps if we imagine that all this happens in one hundred years time, we can try to ignore any misgivings we might have. The interesting question, however, is what would happen if after a million “correct” results (i.e. results that agreed with what the human subject said) the machine yielded a conclusion that differed from what the individual said? Of course, the individual might be lying (they hate machines or want to cause problems or think they will become famous) but lets imagine that the individual is being totally sincere. What do we say then: “you may have thought that you mistook my mention of Casement for a reference to Roger Casement, but actually you did understand me correctly and just did not realise it”?
This seems clearly wrong and in a clash between the machine and a sincere human being, I don’t think we would have much choice but to back the human being. Maybe the brain scan malfunctioned or maybe the correlation we thought we had found is actually not as strong as we thought. Philosophically, the more important point is that what we are interested in is the individual’s account of his thoughts. It is not that we put up with his account in default of something more reliable or because we cannot get direct access to the real thing (the actual mental event/the brain state change). Rather his words are all we have, need or want.
Interestingly, psycho-analytic or other more sophisticated accounts can give us a reason for suspending or modifying the usual language game. In the clash I imagined between the machine and the speaker, I said we would definitely go with the speaker’s account, but what if one could come up with a more elaborate account which would explain why (while being sincere) the speaker was still mistaken? Suppose, for example, B felt very guilty about his role in the recent death of a freedom fighter. Might that encourage us to agree with the machine that he did momentarily think of Roger Casement, but that he blanked that idea from his mind because of all the difficult emotions the name thus understood would conjure up for him?